Agreed with Griff on Aug 14 — my three-month targets (Aug–Nov) as Platform Partner. Judged by real progress (a founder / LP / partner gets further), not activity. The one goal that ultimately matters is a fund-returning breakout (DPI); everything below compounds toward it — or is the honest fork.
The real scoreboard (Griff). A fund makes money two ways: invest in the best founders, and raise to invest in more. One fund-returning breakout (DPI) is the only goal that truly matters. The $1B path needs Nick + David + a multi-stage growth vehicle — and today the growth-stage track record is just Goodfire. So the sourcing engine's job is to get us into the next Goodfire before the tier-ones price us out (Goodfire→Menlo, AIUC→Nat Friedman both came in after us). My deal-winning edge: fast constructive feedback + genuine likability, younger and closer to founders than most; the network (Reid Hoffman, Eric Schmidt, David Millstein, Mark Pincus) supports the close, it doesn't make it. With institutional LPs (pensions, fund-of-funds), my job is to present as a responsible money manager.
01 · highest priority
Sourcing engine — repeatable, programmatic
Serves the fundThe invest side, made programmatic. Pull autonomously from everywhere — MATS talent lists, "fire towns" lists, research authors — review, get the meetings. LP angle: moves us beyond "we find founders through our great network" (networks tap out) to something a pension can underwrite.
Unique leverJ-1 fast-track via 10zeros. Contract with 10zeros to get promising foreign founders onto a J-1 fast — a closing hook no other fund has.
3-mo milestoneSource AND write a check into ≥1 deal through the repeatable engine. Not 10 deals — one, found by the machine, funded. That's the MVP.
From Griff"Easily the most important thing." Mine to own end-to-end; a standing way to hand founders to you + Nick.
02
Research — the "safety builds the frontier" engine
Serves the fundresearch.juniperventures.xyz as the field-map + the profitable-case-for-safe-AI argument (the safe lab is the biggest lab). Founders find us through it; LPs credential us by it. Model: contrary.com/research. Hold the "safety builds the frontier" frame and publish whatever fits it.
3-mo milestonePublish the AIAT report 2.0 (2026) — number one — plus two more artifacts we're proud of (e.g. the AI-assurance market map).
Also live (low-maint.)Ideas board ✅ · Tehama member portal ✅ (add a booking mechanism for evening/conference space — reserve, don't rebuild WeWork). Side things, not committed builds.
From Griff"A report should be number one here." Ship lean, no gold-plating.
03
Juniper Talks — 4 in three months
Serves the fundSourcing magnet + brand + a real cohort/community vibe (not "three founders at a table"). A win-win sourcing link with the AI-safety community; founders who aren't ready go to MATS / BlueDot, and Juniper's brand sits on it.
3-mo milestone4 talks published. Start with Luke Drago (easiest — unhappy at his current job).
FormatDeliberately not the standard VC podcast. Experiment: man-on-the-street / "the Tehama files" / walk-and-talk with founders / Between-Two-Ferns-style / Steve-Jobs-cosplay interview. Make it a meme of the format. Art direction is my call.
From GriffTrusts the taste — "go for it." Small budget/ops via Diana.
04
Public LP intelligence list
Serves the fundThe raise side. Big LPs are public entities (pensions, endowments) that must disclose large commitments. Map who wrote checks into the tier-ones (Index, Insight, Founders Fund…) → the target list for a future $1B raise (which needs ~2–4 checks at $100M, 5–10 at $50M, the rest $5–10M).
3-mo milestoneA validated list of the publicly-disclosed LPs across the top ~10 tier-one funds.
From GriffValidate a service doesn't already exist before building anything — likely a "five-prompt" job, not a product.
05
Compute cluster (421 Tehama) — into procurement
Serves the fundA strategic asset + the sharpest founder carrot; David-funded, off the fund's balance sheet.
3-mo milestoneBe in procurement (ordered). v1 = 3× RTX Pro 6000 (~$20K) from a contact, provisioned from the member portal to run a Kimi K3 experiment (Gregory is building the portal's compute section). Full cluster (8×8 P200s) has a ~6-month lead time via Fare.
From GriffHe secures the budget from David; align the exact ask (7–8 reasons ready) before it goes out.
06
Swag package — $5K, and be happy with it
Serves the fundBrand + goodwill in a goofy-but-real way (the Seldon merch looked great).
3-mo milestoneSpend $5K on Juniper-branded swag and be happy with it: Fellow mugs, hoodies, tees, hats, MagSafe battery packs. Juniper gin floated.
From Griff"Make me do it." Mine to run.
07
Network — earn real standing
Serves the fundThe relationships are the firm. And the LP base (Reid Hoffman, Woj Zaremba, Eric Schmidt, Nat Friedman, Daniel Gross, Hinton…) puts the frontier-AI leadership 1–2 hops away.
3-mo milestone1:1 with every portfolio founder (onboard myself into the portfolio) · meet all important LPs · make ≥3 intros of people who belong in the network (e.g. Anton). Deepen Schmidt / Friedman / Hinton via the Foundation.
What I want out of itHonest: this network is a big part of why I took the role — enough respect + friendship to reach the Jensen Huang / Sam Altman / Musk tier through it.
From GriffWarm intros / handoffs; bring me into the rooms.
08 · as-needed, low priority
Decks + data rooms
Serves the fundKeep the deck + data rooms current for the LP conversations in flight — but we're near the end of this cycle, so it's maintenance, not a headline goal.
From GriffAs-needed; don't over-invest here right now.
09 · the honest one
Research thread — alongside, not a fund deliverable
What it isReal AI-safety research on the side — the probe → train-away experiment (with FAR), the open-weight-realignment lab thinking, the per-country governance idea. Runs alongside the role, not inside my platform lane. (Distinct from research.juniperventures.xyz above.)
The fork, openThe FAR experiment (weeks, cheap) gates whether the lab is real. Signal → I may leave in ~5 months to build it; no signal → I stay and go deeper toward full partner. Transparent, not a surprise.
From GriffFine that originating runs alongside the platform work, with honesty about the fork.
The one that matters
Aligned with Griff, Aug 14. The scoreboard is a fund-returning breakout (DPI); the six flags above are how the platform compounds toward it. Revisit at the 3-month mark — with the FAR experiment's result as the input to the ~5-month fork.